KPI-based Complex Event Evaluation In Discrete Manufacturing Environment

With the rapid economical development, the competition in manufacturing isgetting more and more penetrating. In order to reduce cost and improve performance,many companies extend their information system to shop floor level. Therefore, thereare many information systems which can get the real-time production data. CEP(Complex Event Processing) is an emerging method of processing the real-time data.Using pattern matching, association analysis and other intelligence methods, CEP cantransform the data into complex event. Once complex event is detected, the workshopcan react to it rapidly. Therefore, real-time data provides decision making support toevery level within company. And by doing this, company can improve performance,product quality and reduce cost.Actually, there are too many kinds of complex events in a workshop includingimportant ones and unimportant ones. For this reason, different complex eventsshould have different monitoring weight. This paper aims to create a mathematicalmodeling for calculation of the monitoring weight of complex events. Monitoringweight is measured by monitoring return. The higher the return is, and the larger themonitoring weight is. Monitoring return is computed by five aspects: influence,frequency, predictability, observability and timeliness. Influence can be get accordingto the importance of key performance indicator (KPI), and others are obtained byfuzzy comprehensive evaluation.After reviewing and summarizing of related lieratures, this paper introduces aKPI system by using value chain analysis and key factor analysis. Categorised by thesource of the complex events, a complex event system in workshop is created, and therelationship between the complex events is also researched. After that, themathematical modeling is created. In case study, a physical manufacturing operationcell is redesigned to monitor complex event with RFID technolgy.